Normalised Equi-Angular Recognition Technique of Human Ear Signatures for Use in Biometrics Applications

Human ear is perhaps one of the relatively promising features that can be used in biometrics applications. Hence, the aim of this work is to assess the validity of using human ear recognition in real biometrics applications. A proposed system for ear recognition is presented and experimentally tested. The system employs equiangular ear signatures. Cross Correlation Function (CCF) is employed in the system to find best angular match and apply angular shift to ear-signatures accordingly. Image scene size problem is overcome by using normalised values of the ear-signatures. Re-arrangements of the ear-signature data set is applied using equiangular steps where equally stepped size angles are used to acquire signature data set of same and fixed length array points, hence, comparison and matching of point to point can be readily applicable. Two parameters, namely error energy and Pareto's-based indicator, to assess the signature similarities are proposed and used in the investigation. Results showed the validity of the approach and encourage the adopting of the developed technique in real applications.